Preprint typeset in JHEP style - HYPER VERSION

An Exceptionally Simple Theory of Everything

arXiv:0711.0770v1 [hep-th] 6 Nov 2007

A. Garrett Lisi
SLRI, 722 Tyner Way, Incline Village, NV 89451 E-mail: alisi@hawaii.edu

Abstract: All elds of the standard model and gravity are unied as an E8 principal bundle connection. A non-compact real form of the E8 Lie algebra has G2 and F4 subalgebras which break down to strong su(3), electroweak su(2) x u(1), gravitational so(3,1), the frame-Higgs, and three generations of fermions related by triality. The interactions and dynamics of these 1-form and Grassmann valued parts of an E8 superconnection are described by the curvature and action over a four dimensional base manifold. Keywords: ToE.

Contents
1. Introduction 1.1 A connection with everything 2. The Standard Model Polytope 2.1 Strong G2 2.2 Graviweak F 4 2.2.1 Gravitational D2 2.2.2 Electroweak D2 2.2.3 Graviweak D4 2.2.4 F 4 2.3 F 4 and G2 together 2.4 E8 2.4.1 New particles 2.4.2 E8 triality 3. Dynamics 3.1 Curvature 3.2 Action 3.2.1 Gravity 3.2.2 Other bosons 3.2.3 Fermions 4. Summary 5. Discussion and Conclusion 1 2 4 5 8 8 10 11 13 14 16 21 22 23 23 25 25 26 27 28 28

1. Introduction
We exist in a universe described by mathematics. But which math? Although it is interesting to consider that the universe may be the physical instantiation of all mathematics,[1] there is a classic principle for restricting the possibilities: The mathematics of the universe should be beautiful. A successful description of nature should be a concise, elegant, unied mathematical structure consistent with experience. Hundreds of years of theoretical and experimental work have produced an extremely successful pair of mathematical theories describing our world. The standard model of particles and interactions described by quantum eld theory is a paragon of predictive excellence.

General relativity, a theory of gravity built from pure geometry, is exceedingly elegant and eective in its domain of applicability. Any attempt to describe nature at the foundational level must reproduce these successful theories, and the most sensible course towards unication is to extend them with as little new mathematical machinery as necessary. The further we drift from these experimentally veried foundations, the less likely our mathematics is to correspond with reality. In the absence of new experimental data, we should be very careful, accepting sophisticated mathematical constructions only when they provide a clear simplication. And we should pare and unite existing structures whenever possible. The standard model and general relativity are the best mathematical descriptions we have of our universe. By considering these two theories and following our guiding principles, we will be led to a beautiful unication. 1.1 A connection with everything The building blocks of the standard model and gravity are elds over a four dimensional base manifold. The electroweak and strong gauge elds are described by Lie algebra valued connection 1-forms, W su(2) B u(1) g su(3)

while the gravitational elds are described by the spin connection, so(3, 1) = Cl2 (3, 1) a Cliord bivector valued 1-form, and the frame, e Cl1 (3, 1), a Cliord Figure 1: A periodic table of the standard model. vector valued 1-form. The frame may be combined with a multiplet of Higgs scalar elds, , to interact with the electroweak gauge elds and fermions to give them masses. The fermions are represented as Grassmann valued spinor elds, { e , e, u, . . .}, with the spin connection and gauge elds acting . . . on them in fundamental representations. The electroweak W acts on doublets of left chiral fermions, {[ eL , eL ], . . .}; the strong g acts on triplets of red, green, and blue colored . . r , ug , ub ], . . .}; and the electroweak B acts on all with an interesting pattern of quarks, {[u . . . hypercharges. The left and right chiral parts of the gravitational spin connection, , act on the frame and on the left and right chiral fermions. This structure, depicted in Figure 1, is repeated over three generations of fermions with dierent masses. This diverse collection of elds in various algebras and representations is, inarguably, a mess. It is dicult at rst to believe they can be unied as aspects of a unique mathematical structure but they can. The gauge elds are known to combine naturally as the connection

of a grand unied theory with a larger Lie group, and we continue with unication in this spirit. The spin connection, frame, and Higgs may be viewed as Lie algebra elements and included as parts of a graviweak connection. Relying on the algebraic structure of the exceptional Lie groups, the fermions may also be recast as Lie algebra elements and included naturally as parts of a BRST extended connection.[2, 3] The result of this program is a single principal bundle connection with everything, A = 1 + 1 e + B + W + g + . 2 4 + ( e + e + u + d) + ( + + c + s) + ( + + t + b) . . . . . . . . . . . .

(1.1)

In this connection the bosonic elds, such as the strong g = dxi gi A TA , are Lie algebra valued 1forms, and the fermionic elds, such as u = uA TA , are Lie algebra valued Grassmann numbers. . . (These Grassmann elds may be considered ghosts of former gauge elds, or accepted a priori as parts of this superconnection.) The dynamics are described by the curvature, F = dA + 1 [A, A] . 2 . . (1.2)

with interactions between particles given by their Lie bracket. For example, the interaction between two quarks and a gluon is specied by the Lie bracket between their generators, with a corresponding Feynman vertex,
g g r
g [Vgr , Vug ] = Vur

ug

It is a remarkable property of the exceptional Lie groups that some of their Lie brackets are equivalent to the action of a subgroup on vectors in fundamental representation spaces, just as they occur in the standard model.[4] For example, the bracket between the gluons and a set of colored quarks in A can give the su(3) action on the dening 3, . ur . = g ug . ub .

ur

g, ur + ug + ub . . .

When all standard model particles and interactions are identied this way, the entire ensemble corresponds to a uniquely beautiful Lie group the largest simple exceptional group, E8.

2. The Standard Model Polytope

The structure of a simple Lie algebra is described by its root system. An N dimensional Lie algebra, considered as a vector space, contains an R dimensional subspace, a Cartan subalgebra, spanned by a maximal set of R inter-commuting generators, Ta , [Ta , Tb ] = Ta Tb Tb Ta = 0 1 a, b R

(R is the rank of the Lie algebra) Every element of the Cartan subalgebra, C = C a Ta , acts linearly on the rest of the Lie algebra via the Lie bracket (the adjoint action). The Lie algebra is spanned by the eigenvectors of this action, the root vectors, V , with each corresponding to an eigenvalue, [C, V ] = V =
a

iC a a V

Each of the (N R) non-zero eigenvalues, , (imaginary for real compact groups) is linearly dependent on the coecients of C and corresponds to a point, a root, a , in the space dual to the Cartan subalgebra. The pattern of roots in R dimensions uniquely characterizes the algebra and is independent of the choice of Cartan subalgebra and rotations of the constituent generators. Since the root vectors, V , and Cartan subalgebra generators, Ta , span the Lie algebra, they may be used as convenient generators the Cartan-Weyl basis of the Lie algebra, A = AB TB = Aa Ta + A V The Lie bracket between root vectors corresponds to vector addition between their roots, and to interactions between particles, [V , V ] = V + =

Elements of the Lie algebra and Cartan subalgebra can also act on vectors in the various representation spaces of the group. In these cases the eigenvectors of the Cartan subalgebra (called weight vectors) have eigenvalues corresponding to the generalized roots (called weights) describing the representation. From this more general point of view, the roots are the weights of the Lie algebra elements in the adjoint representation space. Each weight vector, V , corresponds to a type of elementary particle. The R coordinates of each weight are the quantum numbers of the relevant particle with respect to the chosen Cartan subalgebra generators.

(2.1)

G2 g g r g rg g rb g rb gb g g gb qr qr qg qg qb qb

V g3 g8 (T2 iT1 ) 1 0 (T2 iT1 ) 1 0 1/2 3/2 (T5 iT4 ) (T5 iT4 ) 1/2 3/2 (T7 iT6 ) 1/2 3/2 (T7 iT6 ) 1/2 3/2 1/2 1/2 3 [1, 0, 0] 1/2 1/2 3 [1, 0, 0] 1/2 1/2 3 [0, 1, 0] 1/2 1/2 3 [0, 1, 0] [0, 0, 1] 0 1/ 3 [0, 0, 1] 0 1/ 3

g8

g3

Table 1: The su(3) weight vectors and weight coordinates of the gluon, quark, and anti-quark weights form the G2 root system.

2.1 Strong G2 The gluons, g su(3), in the special unitary group of degree three may be represented using the eight Gell-Mann matrices as generators,
i g = g A TA = g A 2 A = C + g V = 3 1 8 i 3 i g 8 g + 3 g g 1 ig 2 g 4 ig 5 g r g rb 2g + 2 3g 1 i i 3 i 8 g gb g rg = 2 g 1 +ig 2 g 3 + 3 g 8 g 6 ig 7 = 2 g + 2 3g 2 i g8 g8 g 4 +ig 5 g 6 +ig 7 g rb g gb 3 3

(2.2)

The Cartan subalgebra, C = g 3 T3 + g 8 T8 , is identied with the diagonal. This gives root vectors particle types corresponding to the six non-zero roots, such as [C, Vgg ] = i g 3 (1/2) + g 8 ( 3/2) Vgg b b Vg g b 0 0 = (T7 iT6 ) = 0 0 0 0 0 1 0
i g gb = g gb Vgg = 2 (g 6 ig 7 )Vgg b b

0 0 = 0

0 0 0 g gb 0 0

for the green anti-blue gluon. (By an abuse of notation, the coecient, such as g gb , has the same label as the particle eigenvector containing the coecient, and as the root the usage is clear from context.) Since the Cartan subalgebra matrix in the standard representation acting on 3, and its dual acting on are diagonal, the weight vectors, V and V , satisfying 3, CV =
a

ig a a V

and

C V = C T V =
a

ig a a V

are the canonical unit vectors of the 3 and The weights for these the su(3) quantum 3. numbers of the quarks and anti-quarks can be read o the diagonals of C and C = C T = C. The set of weights for su(3), the dening 3, and its dual are shown in Table 1. These 3, weights are precisely the 12 roots of the rank two simple exceptional Lie group, G2. The weight vectors and weights of the 3 and are identied as root vectors and roots of G2. The 3 G2 Lie algebra breaks up as g2 = su(3) + 3 + 3 allowing a connection to be separated into the su(3) gluons, g, and the 3 and quarks and 3 anti-quarks, q and q , related by Lie algebra duality. All interactions (2.1) between gluons and . . quarks correspond to vector addition of the roots of G2, such as
g g r
g [Vgr , Vqg ] = Vqr

g g r + q g = q r

1 1 1 (1, 0) + ( 1 , 23 ) = ( 2 , 23 ) 2

qg

We are including these quarks in a simple exceptional Lie algebra, g2, and not merely acting on them with su(3) in some representation. The necessity of specifying a representation for the quarks has been removed a signicant simplication of mathematical structure. And we will see that this simplication does not occur only for the quarks in g2, but for all fermions of the standard model. Just as we represented the gluons in the (3 3) matrix representation (2.2) of su(3), we may choose to represent the gluons and quarks using the smallest irreducible, (7 7), matrix representation of g2,[6] 1 1 1 1 1 qr qr qg qg 1 0 2 q b 2 q b 2 2 2 2 1 b i 8 1 1 qg qr q g 0 g rb g gb 3 2 2 2 qb 0 g8 i 1 1 1 qg qr g rb g gb 2 3 2 2 1 r 1 g 1 i rg q q g rb i g 3 + g 8 (2.3) qb 0 g g +q +q = 2 2 2 2 3 2 1 r i 3 i 1 g qb 0 g 23 g 8 g r 12 q g rb 2 q g 2 2 1 g 1 r gb 1 b i 3 i r g 8 q q g q g 0 2 g + 2 3g 2 2 2
1 qg 2

qr

g gb

1 qr 2

1 qb 2

g rg

i 3 i 8 2g 2 3g

Squaring this matrix gives all interactions between gluons and quarks, equivalent to su(3) acting on quarks and anti-quarks in the fundamental representation spaces. The G2 root system may also be described in three dimensions as the 12 midpoints of the edges of a cube the vertices of a cuboctahedron. These roots are labeled g and qIII in Table 2, with their (x, y, z) coordinates shown. These points may be rotated and scaled, 1 1 1 B2 x 3 1 1 13 3 3 0 (2.4) g = 2 2 y 2 1 1 2 8 g z
6 6 3

G2 + U (1) x

2 B 2 3

g3 1 1 1/2 1/2 1/2 1/2

1/2

g8 0 0 3/2 3/2 3/2 3/2

1/2 3 1/ 3

g g r g rg r g b g rb g gb g gb r qI g qI b qI qr I qg I qb I

1 1 0 1 1 0 1 0 1 1 0 1 0 1 1 0 1 1
1/2 1/2 1/2 1/2 1/2 1/2 1/2 1/2 1/2

0 0 0 0 0 0
1/6

G2 + U (1) x
r qII g qII b qII qr II qg II qb II r qIII g qIII b qIII qr III qg III qb III

y 0 1 0 0 1 0 1 0 1 1 0 1

z 0 0 1 0 0 1 1 1 0 1 1 0

1/6 1/2 1/2 3 1/6 1/6 1/6 1/6 1/2 1/2

1/2 1/2 1/2 1/2 1/2 1/2 1/2 1/2 1/2 1/2 1/2 1/2 1/2 1/2 1/2

1/2 1/2 3 1/2 1/2 3 0 1/ 3

l l

0 0

0 0

1 0 0 1 0 0 0 1 1 0 1 1

2 B 2 g 3 3 1/3 1/2 1/3 1/3 1/3 1/3 1/3 2/3 2/3 2/3 2/3 2/3 2/3 1/2

g8
1/2 3 1/2 3 1/ 3 1/2 3 1/2 3 1/ 3 1/2 3 1/2 3 1/ 3 1/2 3 1/2 3 1/ 3

0 1/2 1/2 0 1/2 1/2 0 1/2 1/2 0

Table 2: Weights of gluons, three series of quarks and anti-quarks, and leptons, in three dimensions, projecting down to the G2 root system in the last two coordinates.

so that dropping the rst, B2 , coordinate gives the projection to the roots in two dimensions. In general, we can nd subalgebras by starting with the root system of a Lie algebra, rotating it until multiple roots match up on parallel lines, and collapsing the root system along these lines to an embedded space of lower dimension a projection. Since the cuboctahedron is the root system of so(6), we have obtained g2 by projecting along a u(1) in the Cartan subalgebra of so(6), so(6) = su(4) = u(1) + su(3) + 3 + u(1) + g2 3 (2.5)

This particular rotation and projection (2.4) generalizes to give the su(n) subalgebra of any so(2n). We can also obtain g2 as a projected subalgebra of so(7) the root system is the so(6) root system plus 6 shorter roots, labeled qII , at the centers of the faces of the cube in the gure of Table 1. The eight weights at the corners of a half-cube, labeled qI and l, also project down to the roots of G2 and the origin, giving leptons and anti-leptons in addition to quarks, su(3) + 3 + + 1 + 3 1 (2.6) These three series of weights in three dimensions, and their rotations into su(3) coordinates, are shown in Table 2. The action of su(3) on quarks and leptons corresponds to its action on these sets of weights, while the u(1)BL quantum number, B2 , is the baryon minus lepton number, related to their hypercharge. The su(3) action does not move fermions between the nine B2 grades in the table each remains in its series, I, II, or III. Since this su(3) and

u(1)BL are commuting subalgebras, our grand unication of gauge elds follows the same path as the Pati-Salam SU (2)L SU (2)R SU (4) GUT.[5] 2.2 Graviweak F 4 The interactions between other gauge elds are more involved and separate from the strong gluons. Most importantly, the weak W acts only on left-chiral fermions, as determined by their gravitational so(3, 1) quantum numbers. Also, the Higgs, , needs to be combined with the gravitational frame, e, to make a 1-form interacting correctly with the electroweak gauge elds and the fermions. These interactions imply that the spin connection, which acts on the frame, and the electroweak gauge elds, which act on the Higgs, must be combined in a graviweak gauge group. The best candidate for this unication is so(7, 1), which breaks up as so(7, 1) = so(3, 1) + so(4) + (4 4) = so(3, 1) + (su(2)L + su(2)R ) + (4 (2 + 2)) (2.7)

and has the desired balance of gravity and left-right symmetric electroweak gauge elds acting on the frame-Higgs. 2.2.1 Gravitational D2 For its action on spinors, gravity is best described using the spacetime Cliord algebra, Cl(3, 1) a Lie algebra with a symmetric product. The four orthonormal Cliord vector generators, 1 = 2 1 2 = 2 2 3 = 2 3 4 = i1 1

are written here as (4 4) Dirac matrices in a chiral representation, built using the Kronecker product of Pauli matrices, 1 = 0 1 1 0 2 = 0 i i 0 3 = 1 0 0 1

These may be used to write the gravitational frame as 0 0 = e T e S 0 0 e S e T e T e S 0 0 e S e T 0 0

e = dxi (ei ) = i

0 4 +e e

e4 e 0

0 eL

eR 0

with left and right chiral parts, eL/R = i(e4 e ), and the coecients, e = ie4 +ie3 T e = ie4 ie3 T e = ie1 e2 S e = ie1 +e2 = e S S

The d2 = so(3, 1) = Cl2 (3, 1) valued gravitational spin connection is written using the six Cliord bivector generators, = 1 [ , ], as 2 = 1 = 2
1 ( 2 i 4 )i

0 ( 1 2
4 +i )i

0 0 ( +i )i S T

= (2.8)

( i )i S T 0

L 0 0 R
1 3 3 3 3 D2G 2i T 1 S 1 L 1 R 2 2 2 L 1 1 1 0 1 1 1 L 0 1 R 1 0 1 R 1 1 0 1 e 0 1 1/2 1/2 S eS 0 1 1/2 1/2 e 1 0 1/2 1/2 T 1 eT 0 1/2 1/2 fL 1/2 1/2 1/2 0 1/2 1/2 1/2 0 fL fR 1/2 1/2 0 1/2 fR 1/2 1/2 0 1/2

with six real coecients redened into the spatial rotation and temporal boost parts, S =
1 2

These relate to the left and right-chiral (selfdual and antiselfdual) parts of the spin connection, L/R = ( )i = S L/R i T

which are sl(2, C) valued but not independent, = . R L The Cartan subalgebra of gravity, in several dierent coordinates, is
3 3 C = 12 12 + 34 34 = S 12 + T 34 =

( 3 i 3 )i3 0 S T 3 +i 3 )i 0 ( S 3 T

3 L i3 0 3 0 R i3

Taking the Lie bracket with C gives root vectors and roots for the spin connection, such as
3 3 1 C, 4 (13 + 14 i23 + i24 ) = i S (2) + 1 T (2) i

Table 3: Gravitational D2 weights for the spin connection, frame, and fermions, in two coordinate systems.

1 4 (13 +14 i23 +i24 )

for L , and weight vectors and weights for the frame, such as i C, 2 (3 4 ) = i 1 3 i T (2) i 2 (3

4 )

for e . The fermions, such as the left-chiral spin-up up quark, u , are in the 4 of the spinor . L T representation space (2.8) with weight vectors, such as [1, 0, 0, 0], equal to the canonical unit vectors, and weights read o the diagonal of C. The collection of elds and their weights are shown in Table 3. The two coordinate systems in the table are related by a rotation and 4 scaling, 1 1 1 3 1 1 3 3 3 L ( + S ) 1 2 i T = 2 1 12 = 12 i 1 T3 (2.9) 3 3 3 R S 2 ( i T + S ) 2 2

Unlike other standard model roots, the roots of so(3, 1) are not all imaginary the 3 coordinates along the T axis are real. The Spin+ (3, 1) Lie group of gravity, with Lie algebra so(3, 1), is neither simple nor compact it is isomorphic to SL(2, C) = SL(2, R) SL(2, R). According to the ADE classication of Lie groups it is still labeled D2 the same as Spin(4) = SU (2) SU (2) since it has the same root system, albeit with one real axis. 2.2.2 Electroweak D2 The electroweak gauge eld, W su(2)L , acts on left-chiral doublets, such as [uL , dL ]. The . . Pati-Salam GUT introduces a partner to this eld, B 1 su(2)R , acting on all right-chiral i fermion doublets. Part of this eld, B 3 2 3 u(1)R , joins with the u(1)BL complement, B 2 , 1 of the strong su(3) to give the electroweak B u(1)Y . The left-right electroweak partner elds may be joined in a d2 partner to gravity, so(4) = su(2)L + su(2)R Since both W and B1 act on the Higgs doublet, [+ , 0 ], it is sensible to consider the 4 real elds of this Higgs doublet to be components of a vector acted on by the so(4). This suggests we proceed as we did for gravity, using a complementary chiral matrix representation for the four orthonormal basis vectors of Cl(4), 1 = 1 1 2 = 1 2 3 = 1 3 4 = 2 1 + 0 Cl1 (4) 0 0

These allow the Higgs vector eld to be written as 0 i4 +

= =

i4 + 0

0 0 = 0

0 1 0 + 0 1 0

with coecients equal to those of the Higgs doublet, + = 1 i2 0 = 3 i4 = 1 + i2 1 = 3 + i4

The d2 = so(4) = Cl2 (4) valued electroweak connection breaks up into two su(2) parts,
wew = 1 wew = 2 i (V +U ) 2 0 U ) i 0 (V 2

i W 2 0 i 0 B1 2

The U and V elds are analogous to the 1 T and S of gravity, and are related to the i electroweak W and B 1 , analogous to the L and R , by the same rotation and scaling 4 (2.9). The Cartan subalgebra, W3 0 0 0 i 0 W 3 0 0 3 3 C = 1 (W 3 + B1 )12 + 1 (W 3 B1 )34 = 3 4 4 2 0 0 B1 0 3 0 0 0 B1

10

gives root vectors and roots for the electroweak elds, such as W , and weight vectors and weights for the Higgs, such as
3 1 C, 2 (3 +i4 ) = i W 3 (1/2)+B1 (1/2) 1 2 (3 +i4 )

W+ B1 B1
W+ + 0 1 L eL R eR uL dL uR dR

3 D2ew W 3 B1

2 B2 3

1 2Y

Q 1 1 1 1 1 1 0 0 0 1 0 1 2/3 1/3 2/3 1/3

for 0 . The fermions are acted on in the standard 4, equivalent to the independent su(2)L and su(2)R action on left and right-chiral Weyl doublets, such as [uL , dL ] and [uR , dR ]. The electroweak D2 weights for various elds are shown in Table 4. The two right-chiral gauge elds, B , are not part 1 of the standard model. They are a necessary part of the Pati-Salam GUT, and presumably have large masses or some other mechanism breaking left-right symmetry and impeding their detection. As in the Pati-Salam GUT, the B2 weights from Table 2 and 3 the B1 weights may be scaled and rotated ((2.4) and (2.9)) into two new coordinates, including the weak hypercharge,
1 2Y 3 = B1 2 B2 3

1 1 0 0 1/2 1/2 1/2 1/2 1/2 1/2 0 0 1/2 1/2 0 0

0 0 1 1 1/2 1/2 1/2 1/2 0 0 1/2 1/2 0 0 1/2 1/2

0 0 0 0 0 0 0 0 1/2 1/2 1/2 1/2 1/6 1/6 1/6 1/6

0 0 1 1 1/2 1/2 1/2 1/2 1/2 1/2 0 1 1/6 1/6 2/3 1/3

Table 4: Weights for electroweak D2, for B2 from Table 2, and electroweak hypercharge and charge.

This scaling implies a weak hypercharge coupling constant of g1 = 3/5 and Weinberg angle satisfying sin2 W = 3/8, typical of almost all grand unied theories. There is also a new quantum number partner to the hypercharge, X, corresponding to the positive combination 3 of quantum numbers B1 and B2 . The hypercharge may be scaled and rotated with W 3 to give the electric charge,
1 Q = W3 + 2Y

These weights, shown in Table 4, are in agreement with the known standard model quantum numbers, and justify our use of the corresponding particle labels. 2.2.3 Graviweak D4 The electroweak d2 = so(4) and gravitational d2 = so(3, 1) combine as commuting parts of a graviweak d4 = so(7, 1). The 4 Higgs elds, , a vector of the electroweak so(4), combine with the 4 gravitational so(3, 1) vectors of the frame, e, into 16 bivector valued elds, e, of the graviweak D4 gauge group. This combination is achieved by adding the weights of Table 3 with those of Table 4 to obtain the weights of D4 in four dimensions, as shown in Table 5. The weights of the fermions also add to give their D4 weights. The fermion weights correspond to the fundamental positive-chiral spinor representation space, 8S+ , of D4. To construct this explicitly, we use Traylings model,[7] and combine

11

our Cl(3, 1) and Cl(4) basis generators into eight Cliord basis vector elements of Cl(7, 1), represented as (16 16) matrices, 1 2 3 4 = = = = 2 3 1 1 2 3 1 2 2 3 1 3 i1 1 1 1 1 2 3 4 = = = = 2 1 1 1 2 1 2 1 2 1 3 1 2 2 1 1 D4
1 3 1 3 2 L 2 R 3 W 3 B1

These allow us to build the spin connection, = 1 , 2 the electroweak connection, wew = 1 , the frame, ew 2 e = e , and the Higgs, = , as Cl(7, 1) valued elds, with the same coecients as before. The frame and Higgs multiply to give the frame-Higgs, e = e , a Cliord bivector valued 1-form. Together, these elds may be written as parts of a Cl2 (7, 1) = so(7, 1) graviweak connection, H 1 = 1 + 1 e + wew 2 4 (2.10)

L / R W B1
/

1 0 0 0 0 1 0 0 0 0 1 0 0 0 0 1
1/2 1/2 1/2 1/2 1/2 1/2 1/2 1/2 1/2 1/2 1/2 1/2 1/2 1/2 1/2 1/2 1/2 1/2 1/2 1/2

0 1/2 0 1/2 The fractions, 1 and 1 , multiplying elds in H 1 are nec2 4 essary for tting gravity and the electroweak connection Table 5: Graviweak D4 roots for 24 together in D4, and for obtaining the correct dynamics bosons and weights for 8S+ fermions. from curvature. The rst 24 weights in Table 5 are the roots of D4. This particular root system has a uniquely beautiful set of symmetries called triality,[8] rotations of the root system by 2 3 that leave it invariant. A triality rotation matrix, T , can permute the coordinates of the root system, 1 3 1 3 3 0001 B1 2 L 2 L 1 3 1 0 0 0 1 3 1 3 2 R 2 L 2 R = 3 = W 0 0 1 0 W3 W3 1 3 3 B3 0100 B1 1 2 R taking each root to its rst triality partner, then to its second, and back satisfying T 3 = 1. As an example, the above triality rotation gives
+ T T T R = T T B1 = T L = R

Since our chosen Cl(7, 1) representation is chiral, H 1 may be represented by its positive-chiral part, the (8 8) rst quadrant of the (16 16) rep, shown here acting on a positive-chiral spinor, 8S+ : 1 i 1 1 3 W+ 4 eR 1 eL 2 L + 2 W 4 e R + 1 i 1 1 3 W 2 L 2 W 4 eR 4 eR 0 eL + i 3 1 1 eR 1 e L 0 B1 4 4 e L + 2 R + 2 B1 1 1 i 3 1 B1 eR 4 e L 4 e L 1 2 R 2 B 1

e T + / e T / e T 0 / e T 1

/ e S + / e S / e S 0 / e S 1 / eL / eL / eR / eR

1/2 1/2 1/2 1/2 1/2 1/2 1/2 1/2 1/2 1/2 1/2 1/2 1/2 1/2

0 0
1/2

1/2 1/2

0 0
1/2

12

showing the equivalence of these roots under this triality rotation. Six of the roots, {W + , W , e + , e 0 , e , e 1 } S S S S are their own triality partners they lie in the plane orthogonal to this triality rotation. The last 8 weights of Table 5, representing one generation of leptons as 8S+ , are rotated by triality into the other fundamental representation spaces of D4: the negative-chiral spinor and the vector, T 8S+ = 8S T 8S = 8V T 8V = 8S+ These two new sets of weights are equivalent to the 8S+ under this triality rotation they carry the same quantum numbers and have the same interactions with the triality rotated roots of D4. Given this relationship, we tentatively consider these three triality partners of 8S+ as the three generations of fermions, such as T e = L L
T = L L T L = e L

The complete set of weights, D4 + (8S+ + 8S + 8V ), including these new triality partners, is the root system of the rank four simple exceptional group, F 4. 2.2.4 F 4 The 48 roots of F 4 are shown in Tables 5 and 6. These roots, in four dimensions, are the vertices of the 24-cell polytope and its dual. Using the breakdown of F 4 into D4 and the three triality-equivalent fundamental representation spaces, f 4 = d4 + (8S+ + 8S + 8V ) = so(7, 1) + (8+8+8) (2.11)

the graviweak bosons (2.10) and three generations of leptons (or quarks) may be written as parts of a F 4 connection, H 1 + ( e + e) + ( + ) + ( + ) . . . . . . Although we are labeling triality partners as fermions of dierent generations, the exact relationship between triality and generations is more complicated and not yet clear to the author. One clue is that the triality partners of F 4 (connected in the gure by pale blue and thin gray lines) may be collapsed to their midpoints to get a g2 subalgebra,
1 3 (1

+ T + T 2 )f 4 = g2 f 4

This triality collapse might relate to a description of graviweak interactions with a group smaller than F 4.[9, 10] It also suggests physical fermions may be linear combinations of triality parners, such as = a fL + b T fL + c T T fL L

13

8S tri L
/

1 3 2 L 3 B1

1 3 3 B3 1 2 R W 1 3 3 1 3 2 L W 2 R 1/2 1/2 0

/ L 0 1/2 1/2 0 / R 1/2 0 0 1/2 / 1 R 0 1/2 /2 0

8V tri L / L / R / R
/

1 3 1 3 2 L 2 R 1 3 3 2 R B1

3 W 3 B1 3 1 3 W 2 L

0 0 1/2 1/2

0 1/2 1/2 0 1/2 1/2 1/2 0 0 1/2 0 0

Table 6: The 8S and 8V weights complete the D4 and 8S+ weight system of Table 5 to form the F 4 root system. The 48 roots are projected from four dimensions to two and plotted, with lines shown between triality partners.

Guided by this triality symmetry, we will continue to label the triality partners with generation labels though this should be understood as an idealization of a more complex and as yet unclear relationship between physical particles and triality partners. If we wished, we could write the constituent particles of F 4 as matrix elements of its smallest irreducible, 26 dimensional representation,[11] as we did for the gluons and quarks (2.3) in G2. We can also compute particle interactions by adding the roots in Tables 5 and 6, such as e + e + = eR L T between an electron, a frame-Higgs, and an electron neutrino. These graviweak interactions, described by the structure of F 4, do not involve anti-fermions or color; to include all standard model interactions we will have to combine F 4 and G2. 2.3 F 4 and G2 together The coordinate axes chosen in Tables 5 and 6 are a good choice for expressing the quantum numbers for gravity and the electroweak elds, but they are not the standard axes for

14

describing the F 4 root system. We can rotate to our other coordinate system,
3 1 3 { 2i T , 1 S , U 3 , V 3 } 2

using a pair of rotations (2.9) and thereby express the 48 roots of F 4 in standard coordi4 nates, shown in Table 7. These coordinate values are described by various permutations of 1, 1/2, and 0; and a similar description of the G2 and U (1) weights from Table 2 is also presented. G2 + U (1) su(3) qI qI l l qII qII qIII qIII perms 1 1 all 1/2 1/2 1/2 two > 0 1/2 1/2 1/2 one > 0 1/2 1/2 1/2 one 1/2 1/2 1/2 one 1 all 1 all 1 1 all 1 1 all x y z # 6 3 3 1 1 3 3 3 3

F4 so(7, 1) 8S+ 8S 8V

1 3 1 3 2i T 2 S

U 3 V 3 perms 1 1 all

# 24

1/2 1/2 1/2 1/2

even# > 0 8 1/2 1/2 1/2 1/2 odd# > 0 8 1 all 8

Table 7: Roots of F 4 and weights of Table 2 described with allowed permutations of coordinate values.

To completely describe every eld in the standard model and gravity we need to combine these two sets of quantum numbers. The graviweak F 4 root system includes the two quantum numbers of so(3, 1) gravity and the two of the su(2)L and su(2)R electroweak elds, with three generations of fermions related through so(7, 1) triality (2.11). The G2 weight system includes the three quantum numbers of the su(3) strong elds and a u(1)BL contributing to hypercharge, with fermions and anti-fermions related through duality (2.6). To match the quantum numbers of all known standard model and gravitational elds, the so(7, 1) of F 4 and su(3) and u(1)BL of G2 + U (1) must act on three generations of 8 fermions for each of the 3 colors of quark, 1 uncolored lepton, and their anti-particles, so(7, 1) + (su(3) + u(1)) + (8 + 8 + 8) (3 + + 1 + 3 1)

(2.12)

as depicted in the periodic table, Figure 1. The weights of these 222 elements corresponding to the quantum numbers of all gravitational and standard model elds exactly match 222 roots out of the 240 of the largest simple exceptional Lie group, E8.

15

2.4 E8 E8 is perhaps the most beautiful structure in all of mathematics, but its very complex. Hermann Nicolai E8 x1 x2 x3 x4 x5 x6 x7 x8 # so(16) 1 1 all perms 112 1/2 ... even# > 0 16S+ 128
Table 8: The 240 roots of E8.

Just as we joined the weights of D2G and D2ew to form the F 4 graviweak root system, the weights of F 4 and G2 may be joined to form the roots of E8 the vertices of the E8 polytope shown in Table 8. Combining these weights in eight dimensions requires the introduction of a new quantum number, w, with values determined by the F 4 and G2 numbers. These quantum numbers uniquely identify each root of E8 as an elementary particle Table 9.

E8

1 3 1 3 2i T 2 S / L W / R B1

U3 V 3 0 1 1 1

w 0 0 0
1/2 1/2

y 0 0 0

F4

G2 #

e+ e e1 e0 eL eL eR eR eL eL eR eR uL dL uR dR uL dL uR dR L L R R L L R R cL sL cR sR cL sL cR sR L L R R L L R R tL bL tR bR b b tL L tR R g x1 x2 x3

1 1 0 1
1/2 1/2 1/2 1/2

D2G 1 4 D2ew 1 4 4 4 1 16 8S+ 8S+ 8S+ 8S 8S 8S 8S 8V 8V 8V 8V 1 1 1 1

1/2

... even#> 0 ... even#> 0 ... even#> 0

1/2 1/2 1/2 1/2 1/2 1/2 1/2 1/2

l l

8 8

... two> 0 8S+

qI 24 qI l l qI qI 1 1 qII qII A2 qII qII qIII 24 8 8 24 24 8 8 24 24 6 6 6 6

... even#> 0 1/2 1/2 ... odd#> 0 1/2 1/2 ... odd#> 0 1/2 1/2 ... odd#> 0 1/2 1/2 ... odd#> 0 1/2 1 1 1 1 1 0 1 0 0 0 0 1 0 1 0 0

... one> 0 1/2 1/2 1/2 1/2 1/2 1/2 ... two> 0 1/2 ... one> 0 0 0 1 1 1 1 1 1 (1 1)
1/2

Table 9: The 240 roots of E8 assigned elementary particle labels according to F 4 and G2 subgroups.

The E8 root system was rst described as a polytope by Thorold Gosset in 1900,[12] and the triacontagonal projection plotted by hand in 1964. This plot,[13] now with elementary particle symbols assigned to their associated roots according to Table 9, is shown in Figure 2, with lines drawn between triality partners.

16

Figure 2: The E8 root system, with each root assigned to an elementary particle eld.

17

The interactions between all standard model and gravitational elds correspond to the Lie brackets between elements of the E8 Lie algebra, and thus to the addition of E8 roots. The Lie algebra breaks into the standard model (2.12) as e8 = f 4 + g2 + 26 7 = (so(7, 1) + (8 + 8 + 8)) + (su(3) + 3 + + (8 + 8 + 8 + 1 + 1) (3 + + 1) 3) 3 = so(7, 1) + (su(3) + u(1)) + (8 + 8 + 8) (3 + + 1 + + u(1) + 3 (3 + 3 1) 3) The 26 is the the traceless exceptional Jordan algebra the smallest irreducible representation space of F 4 and the 7 is the smallest irreducible representation space of G2. Each 8 is the 8S+ , 8S , or 8V of so(7, 1) = d4. And the 3 and are in the dening representation 3 space of su(3) = a2. The last two terms in the last line above represent new particle elds not in the standard model, w u(1) x 3 (3 + 3)

The new x eld carries weak hypercharge and color, has three generations, and couples leptons to quarks. This breakdown of E8 is possible because F 4 is the centralizer of G2 in E8, F 4 = CE8 (G2) To display this subalgebra structure, the E8 root system may be rotated in eight dimensions, projected to two, and plotted, as shown in Figures 3 and 4.1 In these plots, the root coordinates have been transformed by a rotation, 1 1 1 3 1 3 2 2 2i T 2 L 1 3 1 12 1 3 2 R 2 2 S 1 1 3 3 W U 2 2 3 3 1 1 B1 V 2 2 w = w 1 1 1 1 x B2 3 3 3 3 1 1 g 0 y 2 2 g8 z 1 1 2
6 6 3

equivalent to the redenition of the Cartan subalgebra generators according to (2.4) and (2.9). Since the spaces containing the F 4 and G2 root systems are orthogonal in E8, these plots of E8 showing a rotation between the two are especially pretty and convenient for identifying interactions between particles. Also, the central cluster of 72 roots in Figure 4 is the E6 root system, which acts on each of the three colored and anti-colored 27 element clusters of the exceptional Jordan algebra.
1

An animation of this rotation is available at http://deferentialgeometry.org/anim/e8rotation.mov

18

Figure 3: The E8 root system, rotated a little from F 4 towards G2.

19

Figure 4: The E8 root system, rotated a little from G2 towards F 4, showing E6.

20

Inspecting the e6 subalgebra of e8 reveals how the fermions and anti-fermions up to now described as living in real representations are combined in complex representations.[14] The e6 decomposes to graviweak so(7, 1) acting on three complex generations of fermions as e6 = f 4 + (8 + 8 + 8) + u(1) + u(1) 1 = so(7, 1) + (8 + 8 + 8) (1 + + u(1) + u(1) 1) = so(9, 1) + u(1) + 16SC in which the nal u(1) is the complex structure, i, related to the w quantum number, and the 16SC is a complex spinor acted on by the so(9, 1). Although considering its e6, f 4, and g2 subalgebras is useful, the E8 Lie algebra may be broken down to the standard model via a more direct route,[8] e8 = so(7, 1) + so(8) + (8S + 8S + ) + (8S 8S ) + (8V 8V ) = so(7, 1) + (su(3) + u(1) + u(1) + 3 (3 + 3)) + (8 + 8 + 8) (3 + + 1 + 3 1) This decomposition is directly visible in Table 9, in which the rst four coordinate axes are of so(7, 1) and the last four are of so(8). The so(7, 1) decomposes into the graviweak elds, and the so(8) decomposes into strong su(3), u(1)BL , and new elds via the embedding of su(4) in so(8). A matched triality rotation of so(7, 1) and so(8) relates the three generations of fermions. The Lie algebra structure of E8, and its relation to the structure of the standard model, is depicted in Figure 5 a periodic table of E8. A comparison of this structure with Figure 1 shows the extremely close t to the standard model, with only a handful of new particles suggested by the structure of E8. 2.4.1 New particles After all algebraic elements of the standard model have been t to the E8 Lie algebra there are a few e8 elements remaining, representing new, non-standard particles. There are two new quantum numbers, X and w, representing the Pati-Salam partner to weak hypercharge

Figure 5: A periodic table of E8.

21

and a new quantum number related to generations. Each of these corresponds to new u(1) valued elds, X and w, which presumably have large masses impeding their measurement. The use of the Pati-Salam model also implies a non-standard pair of elds, B , interacting 1 with right-chiral fermions. In addition, there is a new eld, x, interacting with leptons and quarks. This eld factors into three generations, x1/2/3 , corresponding to dierent w quantum numbers, and a new Higgs scalar, , for each color and anti-color. The new eld, x, is a joining of x and in the same way e is a joining of the gravitational frame, e, and the Higgs, . Since the frame-Higgs is a composite eld a simple bivector its degrees of freedom do not exhaust the algebraic sector it inhabits. Specically, e = e uses 16 algebraic elements but, because it is simple, has only 4 (for e = e ) plus 4 (for = ) equals 8 algebraic eld degrees of freedom. How or why these 16 algebraic elements are restricted is not understood but this restriction is necessary to recover the standard model and gravity. Because the 18 algebraic degrees of freedom inhabited by x appear amenable to the same sort of factorization as e (see Table 9), it is natural to factor it into three x elds and three colored and three anti-colored Higgs elds, . It could be possible that this new x gives dierent masses to the dierent generations of quarks and leptons, producing the CKM and PMNS matrices. Also, since it mixes leptons and quarks, the existence of this eld predicts proton decay, as does any grand unied theory. The interactions between the new elds, w and x, are analogous to the interactions between the gravitational spin connection and the frame-Higgs, and e. This suggests that a better understanding of the triality relationship between generations will involve how these two sets of elds may be more intimately related. 2.4.2 E8 triality The specic triality matrix chosen to rotate between the fermion generations, in the coordi3 3 3 nates { 1 L , 1 R , W 3 , B1 , w, B2 , g 3 , g 8 }, is 2 2 0 0 0 1 1 0 0 T =
0 0 0 0 1 0 1 0 0
1 2 3 2

3 2 1 2

1 0 0 1

This is a somewhat arbitrary choice, selected for leaving W 3 and color invariant. Once the rst generation of fermions, with correct charges and spins, are assigned to elements of e8, this T rotates them to the second and third generations. The second and third generations only have the correct spins and charges when considered as equivalent under this T . When considered as independent elds with E8 quantum numbers, irrespective of this triality relationship, the second and third generation of elds do not have correct charges and spins. The W 3 and color charges are invariant under our choice of T but the spins and hypercharges are only correct

22

through triality equivalence. This relationship between fermion generations and triality is the least understood aspect of this theory. It is conceivable that there is a more complicated way of assigning three generations of fermions to the E8 roots to get standard model quantum numbers for all three generations without triality equivalence. There is such an assignment known to the author that gives the correct hypercharges for all three generations, but it is not a triality rotation and it produces unusual spins. A correct description of the relationship between triality and generations, if it exists, awaits a better understanding.

3. Dynamics
The dynamics of a connection is specied by the action functional, S[A]. Classically, extrem. izing this action, constrained by boundary data, determines the value of the connection, A(x), . over a region of the base manifold. The value of the connection may also be used to infer topological properties of the base manifold. Quantum mechanically, the action of a connection over the base manifold determines the probability of experiencing that connection.[15] Since quantum mechanics is fundamental to our universe, it may be more direct to describe a set of quantum connections as a spin foam, with states described as a spin network. Under more conventional circumstances, the extensive methods of quantum eld theory for a non-abelian gauge eld may be employed, with propagators and interactions determined by the action. In any case, the dynamics depends on the action, and the action depends on the curvature of the connection. 3.1 Curvature The connection with everything, an e8 valued collection of 1-forms and Grassmann elds, A = H 1 + H 2 + I + II + III e8 . . . . . may be broken up into parts matching the standard model, H 1 = 1 + 1 e + wew so(7, 1) 2 4 so(3, 1) e = (e1 + e2 + e3 + e4 ) (+/0 + /1 ) 4 (2 + 2) wew = W + B 1 su(2)L + su(2)R H 2 = w + B 2 + x + g so(8) w + B2 u(1) + u(1)BL r/g/b + r//) g b x = (x1 + x2 + x3 ) ( 3 (3 + 3) g su(3) I = e + e + u + d . . . . . II = + + c + s . . . . . III = + + t + b . . . . . 8S+ 8S+ 8S 8S 8V 8V (3.1)

23

The curvature of this connection, an e8 valued collection of 2-forms and Grassmann 1-forms,
=

F = dA + 1 [A, A] . 2 . . = dA + AA . . . = F 1 + F 2 + DI + DII + DIII . . . = =

(3.2)

may be computed and broken up into standard model parts. The so(7, 1) part of the curvature, F 1 = F G + F gw + F ew = = = = includes the gravitational so(3, 1) part, the mixed graviweak 4 (2 + part, and the elec2) troweak su(2)L + su(2)R part. The gravitational so(3, 1) part of the curvature is
1 F G = 1 ((d + 1 ) + 8 ee) = 1 (R 1 ee2 ) = 2 2 2 = 8

(3.3)

in which R is the Riemann curvature 2-form, ee is the spacetime area bivector, and 2 is the = amplitude of the Higgs squared. The mixed graviweak 4 (2 + part is 2) T F gw = (de + 1 [, e]) e(d + [W + B 1 , ]) = = eD = 2 (3.4)

in which = is the gravitational torsion and D is the covariant derivative. And the electroweak T su(2)L + su(2)R part of the curvature is F ew = (dW + W W ) + (dB 1 + B 1 B 1 ) = F W + F B1 = = = The so(8) part of the curvature, F 2 = F w + F B2 + F x + F g + xx = = = = = (3.6) (3.5)

includes the u(1) and u(1)BL parts, the mixed 3 (3 + part, and the strong su(3) part. 3) The last term does not easily separate xx contributes to all three parts of F 2 . The = u(1) and u(1)BL parts are F B2 = dB 2 F w = dw = = The mixed 3 (3 + part is 3) F x = (dx+[w+B 2 , x]) x(d+[g, ]) = (Dx)xD = And the strong su(3) part is F g = dg + gg = Due to the exceptional structure of e8, the fermionic part of the curvature for the rst generation is D = d + [H 1 + H 2 , ] . . .
1 = (d + 2 + 1 e) + W L + B 1 R (w + B 2 + x) q g . . . . . 4

with D the covariant massive Dirac derivative in curved spacetime. The second and third fermionic generation parts of this curvature are similar.

24

3.2 Action The most conservative approach to specifying the dynamics is to write down an action agreeing with the known standard model and gravitational action while satisfying our desire for minimalism. With these two motivations in mind, an action for everything can be economically expressed as a modied BF theory action over a four dimensional base manifold, S=

B= + F =

G G G = = 4 B B

B B = =

(3.7)

in which = is the curvature (3.2), B = B + B is an e8 valued collection of 2-form and antiF = = Grassmann 3-form Lagrange multiplier elds, B G is the so(3, 1) part of B, B is the rest of = = = B, = 1 2 3 4 is the Cliord algebra volume element, is the Hodge star, and <> takes = the scalar part (the trace). After varying B and plugging it back in (3.7), this action up to a boundary term is = S=
1 BD + 16G 2 (R 3 2 ) 4 F F e 1 . = = 2

(3.8)

in which is the spacetime volume 4-form, R is the gravitational scalar curvature, and F is e = the non-so(3, 1) part of F1 and F2 . This is recognizable as the action for the standard model = = and gravity, with a cosmological constant related to the Higgs vacuum expectation value, = 3 2 4 The details of the action, and its agreement with the standard model and general relativity, can be worked out for each sector of the E8 Lie algebra. 3.2.1 Gravity The modied BF action for gravity was discovered by MacDowell and Mansouri in 1977, [16] and revived by Smolin, Starodubtsev, and Freidel during their work on loop quantum gravity.[17, 18] The remarkable and surprising fact that gravity, described by the spin connection, , and frame, e, can be described purely in terms of a unied connection, + e, was the seed idea that led to the unication of all elds in a single connection.[2] The gravitational part of the action (3.7) is SG = BGF G + = =
G G G = = 4 B B

in which the gravitational part of the curvature (3.3) is F G = 1 (R 1 ee2 ) so(3, 1) = = 2 = 8 Extremizing the action under variation of the gravitational part of the Lagrange multiplier, B G , requires = 2 1 B G = G F G = G (R 1 ee2 ) = = = 8

25

and plugging this back into the action gives SG =

1 G

F GF G = = =

1 4G

1 (R 1 ee2 )(R 8 ee2 ) = = 8

Multiplying this out gives three terms. The term quadratic in the Riemann curvature is the Chern-Simons boundary term, RR = d (d + 1 ) = = 3 Dropping this, the other two terms give the Palatini action for gravity, SG = =
1 16G 1 16G 1 12

eeee 4 Ree 2 =

3 e 2 R 2 2

equal to the Einstein-Hilbert action with cosmological constant, = 3 2 . The magnitude 4 of the Higgs, 2 , is a conformal factor that can be absorbed into the magnitude of the frame. The vacuum solution to Einsteins equation with positive cosmological constant is de Sitter spacetime (R = ee and R = 4), which should be considered the background vacuum = 6 spacetime for particle interactions in this theory. Since the symmetry of this spacetime is so(4, 1) and not the Poincar group, the Coleman-Mandula theorem does not apply to restrict e the unication of gravity within the larger group. It should be emphasized that the connection (3.1) comprises all elds over the four dimensional base manifold. There are no other elds required to match the elds of the standard model and gravity. The gravitational metric and connection have been supplanted by the frame and spin connection parts of A. The Riemannian geometry of general relativity . has been subsumed by principal bundle geometry a signicant mathematical unication. Devotees of geometry should not despair at this development, as principal bundle geometry is even more natural than Riemannian geometry. A principal bundle with connection can be described purely in terms of a mapping between tangent vector elds (dieomorphisms) on a manifold, without the ab initio introduction of a metric. 3.2.2 Other bosons The part of the action (3.8) for the bosonic, non-so(3, 1) parts of the connection is S =
1 4

F F = =

= Sgw + Sew + S2

in which the relevant parts of the curvature (3.2) are the mixed graviweak part (3.4), the electroweak part (3.5), and the so(8) part (3.6). The mixed graviweak part of the action is Sgw =
1 4

F gw F gw = =

26

= =

1 4 1 4

(= eD) (= eD) T T
=

T T = 2 + (D)e = + T

3 4

(D) (D)

which includes the kinetic Higgs term and gravitational torsion. The electroweak part of the action is 1 Sew = 1 F ew F ew = 1 F W F W 4 F B1 F B1 = = = = 4 = 4 = And the so(8) part of the action is S2 = =
1 4 1 4 1 2

F 2 F 2 = = F w F w = =
1 4

F B2 F B2 = =

1 4

F x F x = =

1 4

F g F g = =

(F w + F B2 + F x + F g + xx) xx = = = =

which includes the action for the gluons and a rst guess at the action for the new elds. This action for the new elds is speculative at this stage and likely to change as our understanding of their role improves. The use of the Hodge dual in this part of the action is required for general covariance but seems somewhat awkward from the viewpoint of this E8 theory. The Hodge star operator requires the frame part, e, to be extracted from the E8 connection, inverted to obtain the coframe, e, and contracted with the curvature. It would be better if there was a natural justication for this procedure, beyond the necessity to agree with known theory. An improved understanding will likely lead to a modication of this part of the action. 3.2.3 Fermions Choosing the anti-Grassmann Lagrange multiplier 3-form to be B = e in the fermionic e part of the action (3.8) gives the massive Dirac action in curved spacetime, Sf = = = = BD .

ee(d + H 1 H 2 ) . . . ee((d + 1 + 1 e + W + B 1 ) (w + B 2 + x + g)) . . 2 4

4 dx |e| (e )i (i + 1 i + Wi + B1i + . . . . 4

+ wi + B2i + xi + gi ) + . . . . . The coframe, e, in this action contracts with the frame part of the graviweak connection, ee = (e )i i dxj (ej ) = (e )i (ei ) = = 4

27

to give the standard Higgs coupling term, . The new, non-standard w and . . x terms are not yet well understood but seem promising for recovering the CKM . matrix. This action works very well for one generation of fermions. The action for the other two generations should be similar, but is related by triality in a way that is not presently understood well enough to write down.

4. Summary
The E8 theory proposed in this work is an exceptionally simple unication of the standard model and gravity. All known elds are parts of an E8 principal bundle connection, A = ( 1 + 1 e + W + B 1 ) + (B 2 + w + x + g) + . 2 4 + ( e + e + u + d) + ( + + c + s) + ( + + t + b) . . . . . . . . . . . . in agreement with the Pati-Salam SU (2)L SU (2)R SU (4) grand unied theory, with a handfull of new elds suggested by the structure of E8. The interactions are described by the curvature of this connection, F = dA + 1 [A, A] . = 2 . . with particle quantum numbers corresponding to the vertices of the E8 polytope in eight dimensions. This structure suggests three fermionic generations related by triality, Te= T= T =e

The action for everything, chosen by hand to be in agreement with the standard model, is concisely expressed as a modied BF theory action, S= B= + B G B G B B F = = = 4= =

with gravity included via the MacDowell-Mansouri technique. The theory has no free parameters. The coupling constants are unied at high energy, and the cosmological constant and masses arise from the vacuum expectation values of the various Higgs elds, g1 =
3/5

g2 = 1

g3 = 1

= 3 2 4

M 0 , 1 , . . .

In sum, everything is described by the pure geometry of an E8 principal bundle, perhaps the most beautiful structure in mathematics.

5. Discussion and Conclusion

There are a remarkable number of coincidences that work exactly right to allow all known 1 elds to be unied as parts of one connection. The factors of 2 and 1 multiplying the spin 4 connection and frame-Higgs result in the correct expressions for the gravitational Riemann

28

curvature and the covariant Dirac derivative in curved spacetime. The fermions t together perfectly in chiral representations under graviweak so(7, 1), and the frame-Higgs has all the correct interactions. This frame-Higgs naturally gets a 4 potential and produces a positive cosmological constant. Finally, and most impressively, the t of all elds of the standard model and gravity to E8 is very tight. The structure of E8 determines exactly the spinor multiplet structure of the known fermions. There are also aspects of this theory that are poorly understood. The relationship between fermion generations and triality is suggested by the structure of E8 but is not perfectly clear a better description may follow from an improved understanding of the new w + x elds and their relation to + e. This relationship may also shed light on how and why nature has chosen a non-compact form, E IX, of E8. Currently, the symmetry breaking and action for the theory are chosen by hand to match the standard model this needs a mathematical justication. Quantum E8 theory follows the methods of quantum eld theory and loop quantum gravity though the details await future work. One enticing possibility is that the gravitational and cosmological constants run from large values at an ultraviolet xed point to the tiny values we encounter at low energies.[19, 20] At the foundational level, a quantum description of the standard model in E8 may be compatible with a spin foam description in terms of braided ribbon networks[21] through the identication of the corresponding nite groups. And there is a more speculative possibility: if the universe is described by an exceptional mathematical structure, this suggests quantum E8 theory may relate to an exceptional Kac-Moody algebra.[22] The theory proposed in this paper represents a comprehensive unication program, describing all elds of the standard model and gravity as parts of a uniquely beautiful mathematical structure. The principal bundle connection and its curvature describe how the E8 manifold twists and turns over spacetime, reproducing all known elds and dynamics through pure geometry. Some aspects of this theory are not yet completely understood, and until they are it should be treated with appropriate skepticism. However, the current match to the standard model and gravity is very good. Future work will either strengthen the correlation to known physics and produce successful predictions for the LHC, or the theory will encounter a fatal contradiction with nature. The lack of extraneous structures and free parameters ensures testable predictions, so it will either succeed or fail spectacularly. If E8 theory is fully successful as a theory of everything, our universe is an exceptionally beautiful shape.

Acknowledgments
The author wishes to thank Peter Woit, Sergei Winitzki, Lee Smolin, Tony Smith, David Richter, Fabrizio Nesti, Sabine Hossenfelder, Laurent Freidel, David Finkelstein, Michael Edwards, James Bjorken, Sundance Bilson-Thompson, John Baez, and Stephon Alexander for valuable discussions and encouragement. Some of the work was carried out under the

29

wonderful hospitality of the Perimeter Institute for Theoretical Physics. This research was supported by grant RFP1-06-07 from The Foundational Questions Institute (fqxi.org).

References
[1] M. Tegmark, Is the theory of everything merely the ultimate ensemble theory? gr-qc/9704009. [2] A.G. Lisi, Cliord bundle formulation of BF gravity generalized to the standard model, gr-qc/0511120. [3] J.W. van Holten, Aspects of BRST Quantization, hep-th/0201124. [4] J.F. Adams, Lectures On Exceptional Lie Groups, University of Chicago (1996). [5] J.C. Pati and A. Salam, Lepton number as the fourth color, Phys. Rev. D 10 (1974) 275. [6] S. Cacciatori et al., Euler angles for G2, hep-th/0503106. [7] G. Trayling, A Geometric Approach to the Standard Model, hep-th/9912231. [8] J. Baez, The Octonions, math.RA/0105155. [9] F. Nesti and R. Percacci, Graviweak Unication, arXiv:0706.3307. [10] S. Alexander, Isogravity: Toward an Electroweak and Gravitational Unication, arXiv:0706.4481. [11] F. Bernardoni et al., Mapping the geometry of the F4 group, arXiv:0705.3978. [12] T. Gosset, On the regular and semi-regular gures in space of n dimensions, Messenger of Mathematics 29 (1900) 43. [13] D. Richter, Triacontagonal coordinates for the E(8) root system, arXiv:0704.3091. [14] F. Grsey, Symmetry Breaking Patterns In E6, Invited talk given at First Workshop on u Grand Unied Theories, Durham, N.H., Apr 10-12 (1980). [15] A.G. Lisi, Quantum mechanics from a universal action reservoir, physics/0605068. [16] S.W. MacDowell and F. Mansouri, Unied Geometric Theory of Gravity and Supergravity, Phys. Rev. Lett. 38 (1977) 739. [17] L. Smolin and A. Starodubtsev, General relativity with a topological phase: An Action principle, hep-th/0311163. [18] L. Freidel and A. Starodubtsev, Quantum gravity in terms of topological observables, hep-th/0501191. [19] M. Reuter, Functional Renormalization Group Equations, Asymptotic Safety, and Quantum Einstein Gravity, arXiv:0708.1317. [20] R. Percacci, Asymptotic Safety, arXiv:0709.3851. [21] S. O. Bilson-Thompson et al., Quantum gravity and the standard model, hep-th/0603022. [22] V.G. Kac, Innite dimensional Lie algebras, Cambridge University Press (1990).

30